AI Sustainable Energy Research Support

The UNDP Sustainable Energy Hub is a network of partners that work alongside countries to transform energy systems though an integrated agenda focused on the policy, technology and financial shifts that shape sustainable economic development. We help countries build net-zero, people-centered societies driven by a just, sustainable energy transition. Our core principle is to promote an integrated agenda that supports energy for development, including by mobilizing partners to enable 500 million additional people to have access to sustainable, reliable, affordable energy by 2025, leaving no-one behind. The global transformation of energy systems has already started, but is being altered by the current geopolitical context. However, this transformation must be accelerated, and it must be done in a way that advances the Sustainable Development Goals. The Sustainable Energy Hub is UNDP’s answer to these challenges. To drive the systems-level change needed, the Sustainable Energy Hub aims to bring about a completely new way of thinking, doing business, connecting people and knowledge. UNDP’s 2022-2025 Strategic Plan has put sustainable energy at the heart of a joint corporate mission, and UN-Energy has pledged to reaching key energy milestones by 2025. UNDP will focus efforts on mobilizing partners and catalzing action to provide access to sustainable, affordable, and reliable energy – both electricity and clean cooking – to 500 million people by 2025, focusing on the world’s poorest communities. UNDP will not do this alone – our role here is to mobilize strong, meaningful, impactful partnerships to deliver action on the ground. We aim to bring about a new way of thinking about energy and advocate for an integrated, inclusive approach, where all stakeholders participate meaningfully in its design.

The Sustainable Energy Academy (SEA) is UNDP’s digital learning platform on sustainable energy for development, covering areas such as energy access, just transition, finance, policy, and data systems. As part of its continued development, we are building an AI-powered research assistant that can help users navigate complex topics and access relevant, high-quality knowledge. A key challenge in developing such systems is not only retrieving documents, but doing so in a way that reflects expert judgment. This exercise is designed to capture how practitioners actually approach research tasks, including how they identify, filter, and prioritize sources under real-world conditions. Your contributions will be used to build a benchmark dataset that reflects expert reasoning. This will guide improvements to retrieval systems, inform how content is structured and surfaced, and support the development of more capable, context-aware AI tools for sustainable energy.

We are seeking online volunteers who are experts in sustainable energy to support improving our Energy AI system, specifically to validate the research engine for retrieving information on sustainable energy for development. Each volunteer will receive a set of 20 research questions, and for each question, you will identify the set of resources you would rely on to construct a strong, well-informed answer, rank them by relevance, and briefly explain why each was selected. You will also document your reasoning and approach, reflecting how you would navigate the research process in practice. This task is designed to capture not only what sources are useful, but how experts think through a research problem, including how they prioritize information, assess credibility, and build an evidence base. The results will be standardize and used to train and energy research model for supporting countries in retrieving relevant publications for informing their energy policy and planning. Key activities include: Reviewing a set of 20 research questions Identifying 5 high-quality resources per question Providing short justifications and rankings for each selection Brief descriptions your research approach and decision-making process Submitting selected resources with standardized metadata